Dreamforce 2025 has concluded for another year and it was full of announcements and product updates. As expected, there was significant focus on the ‘agentic revolution’ and how Salesforce is seeking to empower its customers with intelligent automation.
Thanks to Salesforce+, I was able to catch-up on many of the key sessions. If you’re looking for a quick recap of the major themes, please read-on!
NOTE: Be mindful that Dreamforce includes forward-looking statements. These are subject to change.
'Agentforce 3' is now Agentforce 360
Agentforce 360 connects humans, agents, and data on one trusted platform, helping every employee and every company achieve more than they ever thought possible.
Since Dreamforce 2024, we’ve seen several versions of Agentforce: Agentforce 2, Agentforce 2DX and most recently this year in June, Agentforce 3. It is now used by over 12,000 customers and with each update, the platform’s capabilities and maturity have expanded. At Dreamforce, the latest version was unveiled, Agentforce 360. This is another major step, introducing:
- Agent Builder: A new, streamlined conversational builder designed to simplify building, testing, and refining. The new ‘simulator mode’ also appears to provide an enriched experience.
- Agent Script: A dedicated scripting language that allows teams to define deterministic logic (e.g. conditional logic) whilst retaining flexibility in other areas of the process.
- Agentforce Vibes: Enterprise grade vibe coding (using natural language) for the Salesforce platform, grounded on Salesforce metadata.
- Agentforce Voice: Voice is now generally available and brings human-like voice conversations to Agentforce. Compatible with a number of major Contact Centre as a Service (CCaaS) providers, such as Amazon Connect, Five9 and Vonage, it promises to deliver a seamless experience alongside other channels, such as web, mobile and email.
- Intelligent Context: More sophisticated processing (and chunking) of unstructured data. Supports multi-modal inputs (e.g. images as well as text etc).
The end of the main keynote introduced a new ‘Agentic Enterprise Licensing Agreement’ with unlimited Agentforce and Data Cloud usage. This sits alongside existing pricing models: Flex Credits (pay-per-action), Conversations and PUPM (per user per month for internal use). Whilst details remain limited, Salesforce seems to be exploring predictable pricing structures for its customers. However, for early-stage adopters, Flex Credits will likely remain the most cost-effective route. Work with your account team to establish the best route for you.
Nonetheless, from these updates (and the items below) it is clear that Agentforce is only continuing to gain momentum, with case studies such as PenFed showing how this can be applied and impactful, even in highly regulated environments.
Data Cloud (Data 360)
Data 360 is the new Data Cloud. It’s the same product you have come to love and use, but has a new name, with a ton more innovation coming
Agentforce wasn’t the only product to receive a naming update; Data Cloud has become Data 360. However, the rebrand is not trivial. It reflects Salesforce’s continued investment in the platform and reaffirms the view that it is a core foundation of the agentic enterprise. Specifically, it recognises the importance of unifying fragmented structured and unstructured data to ground agents with trusted context. The overarching goal is to enable agents and businesses to deliver personalised and rich customer experiences.
Alongside the Intelligent Context feature mentioned above, the main updates were:
- Data 360 Clean Rooms: Secure environment for partners to collaborate on shared data without exposing sensitive information; powered by Data 360’s zero-copy pattern.
- Zero Copy Network expansion: An additional 108 Zero Copy connectors are being made available for Data Cloud, including vendors such as IBM and Microsoft.
Salesforce is also simplifying onboarding with a new Setup Agent.
Case studies, such as Meta’s, demonstrate how tailored implementations of Data 360 can yield tangible productivity and efficiency gains (e.g. powering 30+ campaigns per quarter, opening up new market segments and helping to rationalise technology platforms).
The continued expansion of the Zero Copy Network is important. It’s a crucial capability, particularly for large enterprises, that may have existing data lakehouses, who want to avoid operational and regulatory complications of moving data, whilst seeking to enable agentic experiences within Salesforce.
For more information on the latest Data 360 updates, check out this page.
Conversational Interface (Slack)
This is the new Slack, this is Slackbot, this is what AI was meant to be
Slack remained front and centre at Dreamforce 2025, featuring prominently in multiple demos. Positioned as the ‘Agentic OS’, Slack’s integration with Salesforce continues to deepen. However, what stood out to me was its growing role as a hub for cross-platform agent interaction; through the Slack Marketplace. The demo with v0 brought this to life, showing how teams can move from ideation to prototyping and refinement quickly.
The re-imagined Slackbot also received considerable attention, acting as a personal AI assistant. It helps users navigate channels, manage work items and act. When combined with the wider agent ecosystem, it’s clear that Salesforce sees Slack as the unified user experience layer of the agentic era.
Tableau Next: what’s next for CRM Analytics and Tableau?
As you’ve probably gathered, Dreamforce has focused heavily on agentic experiences. You may wonder where analytics fits in? Analytics content highlighted the need to make data actionable to unlock its value.
Over the past year, Tableau Next has gained significant momentum as Salesforce’s API-first, agentic analytics experience. Its semantic layer plays a key role, translating data into business understandable context. This in turn empowers agents and users to draw relevant and trusted insights. Moreover, given Tableau Next is coupled with Data 360, it can help unlock insights from a wide array of business data. Whilst these points are not new, it is important to frame the updates below in this context.
- Open Semantic Interchange (OSI). Salesforce is working with Snowflake, dbt Labs and other industry leaders to define a vendor-neutral open standard for semantic specifications. The OSI aims to ensure consistent business logic across AI and analytics platforms; whilst respecting access controls and privacy.
- Access to Tableau Insights. Tableau insights will soon be available from within Claude and ChatGPT; further empowering analytic visualisations in conversational interfaces.
- No migration required to access Tableau Next. No migration is required to begin using Tableau Next as it will work alongside Tableau Cloud, Desktop, Server and CRMA Products by consuming existing data sources (e.g. Tableau published data sources and CRMA data sets).
These updates open exciting possibilities. OSI seems to represent a major step towards AI and analytics platforms having consistent meanings. This will further help cross-platform collaboration and insights. Moreover, the fact that existing Tableau and CRMA Analytics customers have a path towards Tableau Next without migration lowers the barrier to adoption and furthers the potential to enrich the insights available within Agentforce.
Security
We build security into every layer. We do this so that you can innovate and grow your business, knowing that your data is available and protected. With the evolution of data and privacy regulations, trust and security are more critical than ever
Security has always been a cornerstone of the Salesforce platform, and the shared responsibility model means we all have a role to play in protecting our environments. Recent high-profile attacks have only reinforced this reality; highlighting the increasingly complex and evolving threat landscape we work within.
The Security Keynote brought this to life with a striking demo of Muralidhar Krishnaprasad (Salesforce’s President and CTO of Unified Agentforce Platform) being deep-faked. This was then able to be applied both through telephony and a video call. It’s well worth a watch as it demonstrates the sophistication of today’s threat actors.
From a product specific perspective, there were also several key themes:
- Reminder of the layers of Salesforce’s shared responsibility model. Whilst not ‘new’, it is important to remember the tools available at each layer and who is responsible for them,
- Upcoming enhancements to Security Health Check. Automating runs, opt-in notifications and new risk categories being introduced.
- Agentforce in Security. Further demonstrations of how Agentforce has been embedded into security features (e.g. Security and Privacy center) to make threat detection and remedial actions more efficient.
Overall, Dreamforce 2025 served as a powerful reminder that whilst agentic AI brings incredible potential, it must also be underpinned by strong security foundations. Many of the controls, both soft (e.g. user education) and hard (IP restrictions, MFA and Principle of Least Privilege) remain as relevant as ever. However, from a Salesforce customer or consulting perspective, it is key to ensure that the right mix of configurables and, where needed, enhanceables, are applied to suit your business’s needs (e.g. taking into account the sensitivity of your data, regulatory obligations, etc).
Additional Product Updates
Dreamforce saw a wide array of other updates as well. Some of the other items which caught my attention were as follows:
- Agentforce: In addition to the Agentforce 360 update, Agentforce Grid was a highlight. This capability provides a spreadsheet-style interface to invoke agents in bulk. With various operations possible, this can help to automate processes at scale.
- Field Service: Field Service is receiving significant investment and enhancements, ranging from scheduling optimisation (e.g. see Winter ’26), the ability for field agents to complete forms via voice and improved ‘offline’ capabilities (performance priming).
- Sales: Exploring how agentic experiences could help reimagine the sales cycle (through pipeline, deals and revenue).
- Service: The introduction of Voice opens many opportunities to enrich customer service. Alongside this, ‘Command Centre’ makes it easy to manage interactions, whether by team members of agents, or across channels. The new ‘Agentforce Service Setup’ is also intended to make it easier to adopt Service Cloud capabilities.
- Setup: Updates for Setup are particularly noteworthy. The reimagined Setup experience, powered by Agentforce, empowers teams to perform tasks via natural language (e.g. querying permissions/access, building customisations, etc).
A recurring theme throughout is how agentic experiences are being increasingly integrated throughout the platform.
And more….
Dreamforce had over 1,500 sessions this year, with a number available on Salesforce +. If there are specific topics you’re interested in, I’d recommend going to Salesforce+ to explore more!
True to the Core
True to the core remains one of my favourite sessions at Dreamforce. It’s where the community gets to speak directly with Salesforce senior leadership and product teams. This serves as a barometer of challenges faced by the community, as well as an opportunity to lead to real change.
This year’s True to the Core covered a variety of themes, including:
- Transparency and engagement with the community (e.g. IdeaExchange, True to the Core Deep Dive and Roadmap Exchange)
- Enhancements to Setup (predominately via Agentforce)
- Efficiency gains in Lightning Experience (faster page loads)
- Enhancements to DevOps (e.g. testing capabilities for Agentforce, Flow, etc)
- Salesforce Lightning Design System (SLDS 2) moving into general availability – paving the way for future enhancements, including dark mode!
Beyond these announcements, the discussion touched on broader organisational challenges, such as budget reallocations impacting community programmes. It is also quite telling how it highlighted concerns about the ongoing emphasis on Agentforce, potentially at the expense of other parts of the platform.
If I had to pick one insight, it would be this from Steve Fisher:
In the pre-agentic world...we were more in the performance zone and now we’re kind of in the transformation zone... it’s difficult and uncomfortable but.. I hope that its kind of energising also because we don’t exactly know what the future is going to be, but we’re going to create it together
I found this insight very honest and encouraging. There is no doubt the past few years have brought about enormous change. Not only that, but it has been fast paced. From Salesforce’s perspective, this has been felt at a macro-level (technological change from generative / agentic AI), and also at a micro-level, such as the influence of activist investors. Both have influenced priorities, direction and strategy.
From True to the Core and Dreamforce more widely, there seemed to be a genuine effort to focus on community engagement. Dreamforce saw lots of golden hoodie winners, and reaffirmed the recent relaunch of the architect and Salesforce military programmes. Whilst time will tell how these changes evolve, it is clear the community has a pivotal role to play in navigating and shaping the future of the platform.
Summary & Reflection
So, what were my key takeaways from Dreamforce 2025? Salesforce is clearly doubling down on Agentforce 360 and Data 360. These are both foundational pillars of the ‘agentic enterprise’. This continued evolution will reshape how businesses gain insights, act on data and interact with customers, particularly through conversational interfaces, such as Slack. Agentforce Voice may also refine what is possible for customer service within the platform as well.
Options for customisation are also being reimagined. The new Agentforce setup experience and Agentforce Vibes open new opportunities to democratise customisation. Grounded in underlying Salesforce metadata, this context will help wider audiences explore what is possible. Whilst it is early days, the intended direction of travel is clear, namely, greater accessibility and faster innovation.
Dreamforce 2025 also felt like we’re now firmly in the ‘growth’ stage of Agentforce’s product life cycle. Dreamforce 2024 introduced Agentforce, this year focused more on refinement, scaling and applications, supported by more customer stories and case studies. With new pricing models (e.g. the Agentic Enterprise Licensing Agreement), product enhancements (e.g. new Agent Builder) and related vendor-neutral standards (e.g. Open Semantic Interchange), Salesforce has set the groundwork for interoperability and scalability.
While the reported 12,000 Agentforce customers represents only a small proportion of Salesforce’s total customer base, it is a meaningful milestone. That’s 12,000 customers experimenting, iterating and providing feedback to shape products and commercial models. The pace of change remains intense, but we’re perhaps getting clearer signs of Salesforce’s vision of how the agentic enterprise will look.
Challenges do remain for customers. Aspects such as unifying data, navigating the increasingly complex security landscape and picking the best commercial models remain important considerations. However, Salesforce is providing tools to help customers navigate this and as always, the community will play a key role in maximising its potential. Both from a technology perspective, and also helping to embrace new ways of working. It will be fascinating to see how Agentforce 360 matures, especially as more businesses adopt the solution and it becomes increasingly integral to the platform.
Until next time, thanks for reading! If you found this useful, don’t forget to follow SFDC Penguin on LinkedIn and Twitter/X.
Bonus Penguin Fact 🐧
Following the deep fake of Muralidhar Krishnaprasad mentioned in the Security section, it goes to show that things aren’t always what they seem! This optical illusion (penguin related, of course) may help you determine how you process information. Click here for more info!
Penguins themselves have very adaptable eyesight; being able to retain sharp focus on land and at sea. For example, penguin eyes have adaptations for extremely low light levels which are essential for hunting deep at sea. Learn more here.
Want More?
Found this article useful? Why not share!
Or check out some of our other articles: